Episode 64: Using AI to Advance Scientific Wellness

Dr. Robert Rountree:

This is the Thorne Podcast. The show that navigates the complex world of wellness and explores the latest science behind diet, supplements, and lifestyle approaches to good health. I'm Dr. Robert Rountree, chief medical advisor at Thorne and functional medicine doctor. As a reminder, the recommendations made in this podcast are the recommendations of the individuals who express them and not the recommendations of Thorne statements in this podcast have not been evaluated by the Food and Drug Administration. Any products mentioned are not intended to diagnose, treat, cure, or prevent any disease. Hi, everyone. And welcome back to the Thorne Podcast. This week we're proud to have Dr. Nathan Price on the show again for a second time. Dr. Price is the chief scientific officer at Thorne HealthTech and the co-director of the Hood-Price Lab for systems biomedicine. And I'd actually like to a little bit about what that is. It's great to have you back, Nathan. Maybe you could tell us what that, what's the Hood-Price Lab, what do you do?

Dr. Nathan Price:

Yeah, the Hood-Price Lab is a lab that exists within the Institute for Systems Biology in Seattle. It's where I was full-time before taking the job first at Onegevity and then following the merger at Thorne HealthTech. And the Hood-Price Lab came about four or five, well, maybe five years ago or something like that. Maybe it's a little more than that now. I had a long time first as a mentor, and then a long time partnership with Lee hood. Lee, for those who might not know, is one of the best-known biologists in the world.

Dr. Robert Rountree:

He's a celebrity.

Dr. Nathan Price:

He's a celebrity in the science field for sure. And more than a celebrity, he's someone who has achieved an incredible amount in his career. He was one of the pioneers in molecular immunology, won the Lasker Prize for that, shared with Tonegawa early on and there was some interesting back story because then the Nobel Prize went to Tonegawa and not Lee for some reason. You know the kinds of politic questions about all those kind of things, but anyway, interesting backstory. Lee went on to become the father of this field of systems biology and considered one of the fathers of personalized medicine. Big claim to fame is inventing automated DNA sequencing which made the Human Genome Project possible. And he won the National Medal of Science from President Obama in 2016, I think it was. And so Lee came to me, we'd been doing some work together.

Dr. Nathan Price:

We had co-founded a company together called Arivale and we had just been working closely, and he had become the chief science officer for Providence St. Joseph Health at that time. And he came to me with kind of an interesting proposition and I was a little surprised by it, but he came and said, "I'd really like to merge our lab groups together into one group." He says, "We're doing all these things together. You know the computation, I know the biology obviously." And he thought that with him taking that role and at that stage of his career that this would be a great next step and I ultimately and thankfully ended up saying yes to that. And we formed this integrated lab with about, I think 30 to 40 people, that kind of going up and down over time, biologists, computer scientists, physicists, project managers, software engineers, et cetera, and really working on two big areas, one which is systems medicine.

Dr. Nathan Price:

So how do you take complexities of disease and translate that into something you can do about the perturb networks that are underlying that disease? And then on the opposite side, we both had become really fascinated by the area we call scientific wellness, which is what can you do before a disease initiates often called precision health these days. And so that was this wonderful enterprise. And so now I more moonlight at the lab and spend a little bit of time there now that my full-time job is with Thorne HealthTech but it's a very special place. And we continue to move forward on a number of different fronts. And Lee being who he is is working hard on launching a million-person project to analyze scientific wellness and we'll be involved in that as well. So it's all very exciting.

Dr. Robert Rountree:

Well, I know all of that involves huge amounts of data and that's part of what I want to talk about today is how you manage these massive amounts of data. So let's jump into our discussion which is going to be on artificial intelligence. And a lot of times people hear that AI and they think about R2-D2, Nc3PO, and androids, and the singularity, and when are machines going to be our friends, but really there's a much bigger story to artificial intelligence than that. And then the other term that you hear around that is something called machine learning. So I'm wondering, what does AI really mean when we're talking about healthcare and how are you guys using it at Thorne HealthTech?

Dr. Nathan Price:

Yeah. So let's dive into this and I'm-

Dr. Robert Rountree:

The big topic.

Dr. Nathan Price:

The big topic. Yeah. And I think in terms of how we might go about this is I think we want to have a broad conversation around the kinds of computational models that really pull all of this together in terms of being able to deliver the best information that we can to people. And artificial intelligence, I think it's a really provocative term. It's a really interesting term. It is a term that I always feel like is a bit overused but it's got what everyone wants to talk about. So when we talk about artificial intelligence though, what we're really talking about, and the way I like to talk about it in Thorne HealthTech is health intelligence, but it's really how do you take AI and apply it to health?

Dr. Nathan Price:

So the reason for it is pretty clear. As you start getting into things like genomics, there's no way that a scientist or a physician, even as one with the expertise that you've got, could ever remember how or be able to interpret all three billion base pairs that exist in a person. And of course, they're different person to person. And so, as we think about that, we're essentially forced to put together systems that can capture intelligence, how we think about data, how we apply it, and we have to automate that. And so that has to be done. And that's what we really mean by artificial intelligence within this context. And I want to separate that term from what we're talking about here from the kind of things that we think about in sci-fi because nothing we're doing is remotely [inaudible 00:07:34] anywhere close or anything like that, but is a representation of intelligence that we can apply to data to go to a conclusion.

Dr. Nathan Price:

That's a big element of it. And then machine learning which you brought up, machine learning is primarily about identification of patterns. And so there typically in machine learning, you can do supervised or unsupervised, but typically what you're looking at, especially if it's supervised learning is I have some outcome that I care about. I then want to associate that with a pattern that I can see in accessible data, whether that's a bunch of measures out of your blood or wherever you're pulling them from. And so machine learning is the way that we go about analyzing these huge data sets to identify patterns that are predictive of some health outcome that we care about. And we can get into a whole bunch of examples like that as I'm sure we will, as we go forward, but that's kind of the big picture of it.

Dr. Robert Rountree:

Well, I know we talked in the previous podcast about calorie restrictions and how calorie restriction is known to be beneficial for prolonging healthspan. And so I'm wondering if that might be an example of how we can use AI to figure out what's going on in the body because I've seen presentations on CR that said, "Boy, if we look at the genes that are turned on and off in the body, there's a lot of things happening there." It's not just like you're eating less and one pathway changes in the body, all kinds of genes are activated and all kinds of genes are turned off. And how would you ever be able to interpret that with the small brains that we humans have?

Dr. Nathan Price:

Exactly. And that's where machine learning really shines because you can go in and identify all of these different patterns. And I can give you an example from some of our work, in this case, out of the Hood-Price Lab. And so, one of the things that we looked at from data across thousands of different individuals, and we had microbiomes, and probiomes, and metabolomes, and clinical labs and wearable device anyway, a whole host of data. And we started to look at some of the things that changed with age. And one of the things we found which was quite interesting and we published this in nature and metabolism last year, is that a person's microbiome becomes more unique to them over time if you stay healthy if you age healthily. In other words, you don't get hospitalized, you don't get on a bunch of drugs.

Dr. Nathan Price:

If you stay healthy, starting about 40 or 50, it ticks up. And by uniqueness, what I mean is that your microbiome looks less like anyone else's microbiome. And that kind of goes up over time. And so you can analyze these kind of things in the big data. It's not something we know, it's something we machine learn. It just comes out of the data. And then you can ask questions like, "Well, what kind of metabolites do we see in the blood that come from the microbiome? Are there commonalities that are found as you go through that process?"

Dr. Nathan Price:

And it turned out that even as people's microbiomes were diverging and becoming less similar there was unified metabolic processes that were getting turned on that were associated with those who had stayed healthy versus not. In fact, they were so strong that we could predict. I had predictive power for in an elderly cohort of who would live for the next four years versus not. Wow. And so you could actually see those kind of things. So I'm not trying to go down that rabbit hole right now but just that's an example of, right? We have a lot of data, we set it up, looking at different cohorts that we know something interesting is going to happen with, something important. And then you just let the data tell you what is associated. Then the whole issue and what really comes out is then sorting out what is correlative from what is causal and that's where the real challenge, and.

Dr. Robert Rountree:

There's the rub. I have to say that I worked my way through college in a microlab plating cultures and using different media and growing a handful of different bacteria. And that's what we knew about that got microbiome back then is if somebody had a problem, if they had diarrhea or inflammation of some sort, we'd try to find out if there were parasites or bugs. Again, it was a handful of things that we looked for. And if they weren't there, we'd say, "Well, you don't have salmonella. You don't have Shigella. I don't know what it is. Hopefully, it'll get better." And now we do this DNA sequencing of microbes and we don't even call them microbes. We call them operational taxonomic units, right?

Dr. Nathan Price:

Right.

Dr. Robert Rountree:

Because we're, Well, those aren't really particular bacteria, they're just strands of DNA. And what I'm saying is that without artificial intelligence, I don't know how we'd be able to make sense of that data because there's so much information that comes out when you do a fecal sample.

Dr. Nathan Price:

Exactly. And that's really true because the amount of information just becomes overwhelming.

Dr. Robert Rountree:

Overwhelming.

Dr. Nathan Price:

And so it is a challenge for all of us in this field of how you take all that information and you separate signal from noise, which is really the name of the game. And it's interesting because our brains, we have to do that all the time, right? We ignore the vast amount of signal that's around us. Just think of something super mundane like in the old days people used to drive from work to home or something like that. So if people could remember that, let's say you have to do something like that and you're driving home and we find that to be a pretty simple task most of the time. But the amount of total information around is massive. Where is every molecule of air going? How is every blade of grass moving? You get the point. There's an incredible, massive amount of information around us all the time and yet we navigate perfectly well because we understand what matters and the information reduction that we do is massive, right?

Dr. Nathan Price:

We understand that all that really matters, right? The road is clear and that no person is crossing or that we know it's a light that's red. We recognize those signals. And so what we're trying to build in biology is in areas where we lack intuition trying to figure out, "Okay, how do I take this massive amount of data and understand the 99.99% of it that I don't actually need? How do I chunk it into the stuff that I actually do need?" And that's what all these efforts are really about is figuring out what is it that matters and how do I chunk information so that when I see this huge mass of changes in molecules and I actually understand what is happening in a way that I can be intelligent about how I interface in that. And you need AI as an interface between that massive amount of data and something you can intuit and think about.

Dr. Robert Rountree:

So that brings us to one potential clinical application that I know Thorne is involved in, and that's digital twins. And I'm wondering if, can we talk about that? Is that a secret or is that-

Dr. Nathan Price:

Yeah, we can talk about something.

Dr. Robert Rountree:

.... that we can touch on.

Dr. Nathan Price:

Yeah. I can at least touch on it. Something I'm really excited about even though maybe I can't get into all the details on it.

Dr. Robert Rountree:

Well, maybe just define it and how it might be useful.

Dr. Nathan Price:

So this is one of the areas that I'm most excited about and there have been the rise of digital twins recently to prominence and what a digital twin means is it is a representation of your body, your physiology in a computer. We've partnered with a company called EmbodyBio that's led by Tom Peterson who's the CEO there and someone that I've known and highly respected for probably over a decade now. And we got together in detail a couple of years ago, and then we've entered into an exclusive partnership with Thorne and EmbodyBio for brain health in the wellness space. And so what we've focused on there is to build a dynamic model that simulates how does your brain maintain health? And so this is really a scientific, what I like to call a scientific wellness point of view on attacking a disease which is you don't start with all the things that you know about the disease.

Dr. Nathan Price:

Start with building a model of what the brain has to do to stay alive. And so what we've done there with Embody is to model those processes. And then what you can do is you can then take measurements from a person and you can contextualize what does it mean when you have different types of genetics, kind of blood measures, different kinds of exercise behaviors, different kinds of on, and on, and on. And we've simulated 10 and a half million digital twins of what happens under all these different conditions. And we've compared it against a mountain of data related to things like Alzheimer's disease. And you can recapitulate the fraction of people who get Alzheimer's disease at every age for every genotype, for example, for APOE. And you can get that really quite accurately. You can simulate the known effects of statins and certain statins that are beneficial and those that are probably not beneficial.

Dr. Robert Rountree:

For a particular individual.

Dr. Nathan Price:

For a particular individual. As you get a distribution, you get a readout and which kind of people based on measurements would be likely to benefit or not. And we can go through that and we've compared it now against just a mountain of data. Basically all the published clinical trials, it quantitatively looks at the data that comes out of positive trials like the finger study. It recapitulates why things like BACE 1 inhibitors were harmful to patients and why that is, and essentially and without getting too much into it but what it really shows is one and this is a big topic in the field, but if you ask people what causes Alzheimer's disease the answer, that they're caused by these amyloid plaques.

Dr. Robert Rountree:

Build up of amyloid causes inflammation, end of story.

Dr. Nathan Price:

Yeah. And we think-

Dr. Robert Rountree:

Get rid of the amyloid and you're fine.

Dr. Nathan Price:

Right.

Dr. Robert Rountree:

Except it doesn't work.

Dr. Nathan Price:

Except it doesn't work. We think that is absolutely wrong. So the elements there is because we have done 400 clinical trials targeting amyloid and in fact, even the drug that was approved last year, right? Why was it approved? Well, it clears amyloid but doesn't help cognition. So at its core, we're really convinced that Alzheimer's is primarily a metabolic disease. What I mean by that is if you just think about some very basic facts, so your brain 20% of the energy in your body, and it's about 2% of your body mass. So it is an energy hog sitting up here on the top of your body here. And if you don't have enough energy then and you go into negative energy balance then certain of your neurons will die.

Dr. Robert Rountree:

They basically get fried.

Dr. Nathan Price:

Yeah. They can't sustain. So one of the things that happens as you get older is your ability to perfuse oxygen in your brain goes down. So you get worse. And as you lose that ability to have oxygen, it becomes harder for you to generate as much energy in your brain as you used to. And there are certain areas in your brain which are less perfused where it's harder to get at. And there's a really fascinating study that came out even just last year that had pet scans of this. So in these regions, you go into what, hypometabolism. Basically, you go into negative energy balance. You're not able to create enough and those neurons die. So when that happens, then the total amount of synapse firing you have in your brain goes down. And when that happens, your brain has to do something to recover in order to keep your cognition going.

Dr. Nathan Price:

So what it actually has to do is you need your background firing. There's always a threshold and you need that background firing to go. And so when you're doing that, you need to secrete a molecule then that will have an impact on increasing that background noise such that you can keep firing because this is all based on something called [inaudible 00:20:28] learning which is what fires together wires together. It's how the brain basically learns. And what is that substance that the brain secretes to increase that background noise so that the total synapse firing goes up? It's called beta-amyloid.

Dr. Robert Rountree:

Beta-amyloid. Yeah.

Dr. Nathan Price:

Beta-amyloid does that. So-

Dr. Robert Rountree:

It's not the cause, it's the effect.

Dr. Nathan Price:

It is the compensatory mechanism. So you are literally attacking the compensatory mechanism. So it's a great biomarker. It is a biomarker of a problem, but it is a lousy, lousy target.

Dr. Robert Rountree:

It's not the thing where we need to go after.

Dr. Nathan Price:

Not the thing. Now it does get complicated because this increased synapse firing comes with an energy cost and that energy cost in fact does have it pushes more pressure on because you are exuding energy. So there is an indirect effect back which I think is why it's been so hard to sort out. But anyway, we've built this model. It uses data from 800 different papers including 40% of them that aren't even Alzheimer's papers but they're just on how the brain is wired. So we can now simulate all of this and we can look at the impact of different style of intervention and we can predict it based on mechanistic models of what it's likely to do. And one of the big things from a product perspective is that it in fact predicts that a compound becomes rate-limiting as you're going through these, well, I guess I didn't get into this.

Dr. Nathan Price:

As you're trying to keep this energy up the other big fact is that these cholesterol recycle pathways become really important. That's where APOE comes in and is really crucial. And in fact, we can simulate the rate differences in cholesterol trafficking from APOE four, three, and two. And just putting that plus oxygen perfusion loss as you get older, just those two facts will recapitulate for you the Kaplan-Meier Curves on when people get Alzheimer's at what rate for all those the different APOE genotypes, you can just predict that. The other element about this kind of approach is it tells you immediately why exercise is... As long as you're thinking about amyloid, it's not a direct line. It's like, "Okay, well, all right, what does exercise do to amyloid?" Exercise just keeps your oxygen perfusion good which keeps you in positive energy balance and you don't.

Dr. Nathan Price:

Anyway, we could go down this rabbit hole a lot, and we've done this for two years. So we have a massive amount of information about this now. And it integrates just data across many, many studies from lots of great researchers out there in the Alzheimer's space. But that's the kind of thing you can do. You can build digital twins, you can build these complex models of what's happening in the maintaining wellness. And then you can use that to then go in. There's a whole bunch of things we could get into like that but it's just there's a huge power in using computational models to go from thinking about lots of individual papers to a unified model that has to explain all of the data at the same time and that's hard to do.

Dr. Robert Rountree:

And we couldn't do it without computers. I mean, that's the whole artificial thing as there's just so much data, we couldn't come up with these conclusions without a little help from our mechanical friends. So this has got potential to affect all branches of medicine, right? Neurology, gastroenterology, immunology. There's just so many things we can do with this.

Dr. Nathan Price:

Absolutely. You can build these out around all those areas, every known mechanism of aging, et cetera, et cetera. That's where we really, yeah.

Dr. Robert Rountree:

All right. We needed to take a break. We'll come back and answer some questions from our listeners afterwards. Are you ready to take the guesswork out of good health? If you are, then Thorne makes it easy with simple health tests that offer deep insights into what's going on inside your body. Choose for multiple tests that analyze for sleep, stress, weight management, biological age, the gut microbiome, and more. Thornes at home health tests measure your personal biomarkers, providing detailed insights that help you identify potential health risk and specific areas of improvement, plus each one provides individualized recommendations for diet, exercise, and supplementation. Visit thorne.com to learn more about Thorne's health desk and to start your new health discoveries today. That's T-H-O-R-N-E.com. And we're back. So now it's time to answer some questions from the community. We always appreciate you sending these questions in. Our first question this week comes from a listener who asks, "How far away are we from AI replacing physicians? Do I need to worry that all this is going to put me out of a job, Nathan?"

Dr. Nathan Price:

Yeah, I think we're a long way from that. AI is not going to replace physicians, but physicians who use AI will replace physicians who don't. And I think that's really a good way to put it because it is an essential tool. And there are certain areas where AI will become really important quickly things like interpreting imaging data in certain cases AI can [crosstalk 00:26:34].

Dr. Robert Rountree:

Yeah. Mammograms.

Dr. Nathan Price:

Yeah, because you can just train it on a lot of information. It gets good at it. It's good at classifying. So AI is really good at certain tasks but one of the things that I think is really important for people to understand, and it does come back to what I alluded to at the beginning which is that I think AI is often an overused term and misinterpreted because the kinds of things that computers are incredibly good at are not the same kind of things that humans are really good at. So you can take a human child, you show them a cat and a dog. They see five of them and they can tell the difference between cats and dogs forever.

Dr. Nathan Price:

When we tried to do that with deep learning it took 180 level deep preceptor on in 50 million examples for a computer to actually get good at. For a fact, for a decade we couldn't get a computer that was any good at telling the difference between a dog and a cat. And so it's just wired differently. And yet, if you want multiply numbers that are massive or analyze huge data sets we can't even possibly do it. So then the likely the evolution of this, at least in the shorter term for sure is the notion of humans being augmented in their abilities by partnering with AIs that are good at the kind of things we're not good at.

Dr. Robert Rountree:

Yeah. I mean, we're already using our cellphones all day or our laptops, it's augmented reality. I've got a pretty good size medical library in my office that I almost never use, right? But [crosstalk 00:28:01] on your phone.

Dr. Nathan Price:

Yeah, exactly.

Dr. Robert Rountree:

It's like I got one them my phone that where I can look things up. Yeah. This is already happening. So what ways is Thorne looking at expanding into AI, what's the future for Thorne?

Dr. Nathan Price:

Yeah. So there's a number of different ways that we're going at this. One is we are working on expanding the range of testing that we're doing. So we'll be able in a cost-effective way to make a lot larger numbers of measurements. And the idea behind the AI is then to be able to deliver really personalized insights to the individual based on these kind of molecular profiles all done under HIPAA and all done in ways that are private for the individual but we'll give a lot more insight. Another is that we're building these deep knowledge graphs that represent what we know from the literature, from intervention studies, and so forth about what the various products in the Thorne line will do and then how they fact systems in the body which will be deeply informed by these denser measurements that we will be making as well.

Dr. Nathan Price:

And so as you get into that, then what we'll be doing is building out further our learning system so that we can identify when someone has an issue, when they take the products when they retest, and then we have a learning system that then finds out, "Okay, did what we recommend, do for the person what they were hoping it would." And when it does, we see that and we record it and if it doesn't then we learn from that so that we're better the next time, right? That doesn't work on this individual's digital twin like it did on the others, why? And we'll get better and better at figuring out those kind of things. We're also very interested in connecting in with physicians who are driving forward in their practices and utilizing some of the products and solutions that we have so that we can deliver what we call health intelligence as a service.

Dr. Nathan Price:

But basically what this is making available to physicians or other companies where appropriate where we can actually give access to what we understand about what drives wellness in these AI representations that we're building from what is known in the scientific literature. And then we'll augment that also with this hybrid of not only what we're pulling out of the literature but also what we're doing in the machine learning so that we're fusing together both what you might think of as knowledge-based AI and data-based AI, bringing those together so that we can give as deeper insights as possible. And that's really essential to where we got to go in the future.

Dr. Robert Rountree:

I know. On a previous podcast, I talked with Dr. Mary Kay Ross also from Thorne about our brain health program. And it seems like AI is kind of a perfect fit for that because someone say complains a brain fog or forgetfulness, they're not diagnosed with Alzheimer's disease or dementia but something isn't quite right. We need better tools for assessing where their brain is. And we need better tools for determining whether something we're giving them like phosphatidylcholine, you mentioned is having an impact. I know in clinical practice, if I say, "Okay, you're forgetful, take the supplement." And then they come back in two months, "How's it working?" "Well, I think it's helping." It'd be nice to have more specific metrics that we could use so that I could say, "See, this is how you've changed in your assessment of, here's how your brain is functioning differently. Here's the metabolic markers we've looked at that have shown improvements."

Dr. Nathan Price:

And that's exactly right. In fact, the brain health is where we're really driving a lot of this in its first iteration or at least the next, I should say the next round, the next upgrade iteration. And that's exactly right because we can fold together modules that represent how do you take a set of measurements to make precise recommendations. We're working hard now to fuse that in with what we've learned from these digital twin models so that we can actually represent for an individual.

Dr. Nathan Price:

But if we know a little bit about genetics and blood work and so forth, how to say well, and also cognitive testing that we're looking at that that would enable us to then be go in, make recommendations that we think are going to be impactful for the person and have before and after tests so that they can see for themselves. "Okay. Did it have the effect that was hoped for?" And on the back end, we can also get an assessment of whether or not our recommendations are working for people as at a highest rate as we'd like and iterate and get better and better at that. So that [crosstalk 00:33:04].

Dr. Robert Rountree:

Otherwise, it's all subjective.

Dr. Nathan Price:

That's right.

Dr. Robert Rountree:

It's totally subjective.

Dr. Nathan Price:

It's subjective. And that's the other thing we're really trying to do towards the AI which is to make more visible, more quantitative, the kinds of changes that people are seeing so that it is obvious and clear where the impact is coming.

Dr. Robert Rountree:

How much blood or stool samples or whatever and data do we need to construct this digital twin? Do we need... People always complain when I order labs and they say, "Well, they took four vials. Oh my God." Is this like a massive amount of data you need or is this a fairly straightforward thing?

Dr. Nathan Price:

So I'm going to treat that as a couple of different questions. So first I'll answer the exact question which they're asking which is how much blood do you need. And in the long run, I don't think it's going to be very much. Without getting too specific on what we're working on I'll just say that a lot of multiomics technologies allow you to make thousands of measures from very small amounts of blood. And we can think about processes there that aren't based around the old standard clinical labs because a lot of people have tried to miniaturize those to get into that and people know how that's failed but in the science realm, we've already developed tests that work on very small amounts of blood and you can get thousands of measurements out of there and those already work. Now, the question is, is there enough information in there for you to say something really relevant to health?

Dr. Nathan Price:

And we are working on, without getting into details, we're working on some very significant partnerships that will accelerate that in a major way. And so we're really, really interested in driving that forward so that we can do a lot from even an at home painless, a blood draw experience that's easy for people. So that's where we're really going. And then the question is how much information we can get off of that so that we can construct as good... A digital twin will never be perfect but that we're really doing a lot of work right now to evaluate what are the most important kinds of information to get and then we'll be deploying that to see how well we can do that and how much of a difference it'll make for people.

Dr. Robert Rountree:

And it sounds like you can get a fair amount of information from that single blood draw. You don't have to do daily blood draws or stool samples. And I harken back to, I don't remember his first name, Dr. Snyder at Stanford that-

Dr. Nathan Price:

Mike Snyder [crosstalk 00:35:44].

Dr. Robert Rountree:

.... measured everything for a year.

Dr. Nathan Price:

Oh yeah. Even more than that. Yeah. Mike's been a real pioneer in that space and yeah, I interfaced with him a fair bit over the years. Yeah. He's [crosstalk 00:35:55].

Dr. Robert Rountree:

So a single sample can tell you a lot.

Dr. Nathan Price:

Absolutely. You can get a lot from it. Having dynamic measures is the best. It's not hard from these technologies to make tens of thousands of measurements from a small amount of blood.

Dr. Robert Rountree:

So I think the one last question I would ask is a mixture of several questions which is should we be worried about AI? Is the data safe? Do we worry about leaks? This has been around since they started doing genetic testing. I remember when the 23andMe first came out, people said, "Oh my God, this company's going to know all about my genetic risk for disease and that data's going to get out there and I won't be able to get insurance." So this kind of started with that genetic variant testing but I think the question has continued. "You're putting all my health data into a computer. How safe is it?"

Dr. Nathan Price:

Yeah. And that's a very fair question. At Thorne HealthTech, I'll just say we take this very seriously. We just actually went through an external review of all of our computation, all our approaches for HIPAA, for data privacy. So we had an external group, they come, they intentionally try to break in to see if they can and we just got a top rating from that group. So with no problem, no red flag. So, that we take very seriously. The other thing that I will say though is that there is a huge benefit to choosing to share health data. And I just want to make a little bit of a comment about that because in other domains of our lives, we have given away an incredible amount of information, obviously to the big tech companies to finance companies and so forth in order to get the returns that we wanted in various different ways.

Dr. Nathan Price:

And there's of course, lots of social battles going on over that right now. The impact in health data and I had this question that was asked to me quite a long time ago, maybe geez, maybe almost 20 years ago now, 18 years ago but it stuck with me a long time which someone was asking me in the very early days, "Well, what's the value of a genome?" And it's actually very interesting because the value of a genome is very little because you don't know how to interpret it. You don't know what it means. But as the number of genomes gets larger where we have associated health information, this is true for the other illness as well. But as you have more associated health information, then every time we share that, then the value of every person's genome to them goes up. And so it's kind of a really interesting case because if you think of something like gold, why is gold valuable?

Dr. Nathan Price:

Because it's rare. But genomes become more valuable the less rare they are. So it's kind of an incredible amount of shared information. And when I say valuable, what I mean by that is we unlock the secrets of what drives cancers in early life? What drives Alzheimer's disease? What drives cardiovascular disease? Why do children die too young? It's et cetera, et cetera, et cetera, et cetera. And so, even though we have a long way down these paths, the collective benefit from doing this is huge. And so the way that I think about this, we can't do that if we don't share this kind of information, and that's just a fundamental property of the universe, right?

Dr. Nathan Price:

Scientifically it's just impossible if we don't. And the other side is how do we not cause problems for people from a legal and social standpoint? That's our own behavior on ourselves. So I don't know. This is just me from the kind of perspective that I have, but I would like to believe that as a society, we can deal with something that is a fundamental truth which is that we have to put this data together and understand how it affects health if we're going to conquer all of these issues and that we can somehow figure out how to have good legal and social structures so that we don't use this data against ourselves. And I'd like to believe we're capable of that, so.

Dr. Robert Rountree:

And I'd like to believe that as well.

Dr. Nathan Price:

[crosstalk 00:40:07] I'd like to see us go, we'll see.

Dr. Robert Rountree:

And so I also vote for crowdsourcing information and putting that out. I every now and then I get a letter and email from 23andMe that says, "Just got a question for you? Do you like bitter foods?" Right. Something really basic just to help them confirm, well, a hypothesis they have. If you got this gene, you're more likely to enjoy bitter foods or not enjoy bitter foods. And so these seemingly innocuous kind of questions that they ask can lead to really profound discoveries about how we function as humans.

Dr. Nathan Price:

Yeah. And maybe if I could just add one more thing because there are a lot of movements out there right now. And I think that's really good for patients' rights, individuals' rights of often that you own your own data that you can ask to have data removed if you are worried about it, that you always have to give consent. We never take anything without consent and et cetera, et cetera. So all those things are important and in place and part of that structure that I think we can do to do this in the right way where people benefit.

Dr. Robert Rountree:

Well, all right folks, that's all the time we have this week. Nathan Dr. Price, thank you so much for coming back on the podcast. I've asked you this before, but just quickly, where can people go if they want to follow your work?

Dr. Nathan Price:

Yeah. So they can follow me on LinkedIn. Just Nathan Price Thorne will pop that up. Or I'm also on Twitter at the handle is at @ISBNathanPrice.

Dr. Robert Rountree:

So that was Dr. Nathan Price, the chief scientific officer at Thorne HealthTech. Artificial intelligence and its role in healthcare. Maybe I should say health intelligence because I think that's a term that we're favoring, health intelligence.

Dr. Nathan Price:

Yeah. A few people are using that. Yeah.

Dr. Robert Rountree:

Yes. We'll keep using health intelligence. As always thank you, everyone for listening and we look forward to talking with you in the future. Thanks for listening to the Thorne Podcast. Make sure to never miss an episode by subscribing to the show on your podcast app of choice. If you've got a health or wellness question you'd like answered, simply follow our Instagram and shoot a message to @ThorneHealth. You can also learn more about the topics we discussed by visiting thorne.com and checking out the latest news videos and stories on Thorne's Take 5 Daily Blog. Once again, thanks for tuning in, and don't forget to join us next time for another episode of the Thorne Podcast.